EP0126686A1 - Optical waveguide employing a diffraction grating - Google Patents

Abstract

This structure includes a single mode input guide (20) and two output guides (21, 22). The structure of the invention is characterized in that it further comprises a diffraction grating (24) disposed at the junction between the inlet guide and the three outlet guides, this grating having a pitch which defines only two directions diffraction of order +1 and -1 which coincide with the directions of the two output guides. The direction of order 0 coincides with a third exit guide. Application in integrated optics.

Description

The present invention relates to an optical guide structure. It finds an application in integrated optics.

The structure of the invention is similar to a device known as a "Y junction" which is shown diagrammatically in FIG. 1. Such a device comprises an optical input waveguide 10, of direction De, extended by two guides of output 11 and 12, of directions Ds ₁ , _D s ₂ inclined respectively by + a and -α relative to the direction De of the input guide. The structure also includes a flared zone 13 forming the junction between the inlet guide and the outlet guides. An optical beam entering the junction through the inlet guide is divided into two beams which propagate in the two outlet guides.

To be able to better specify the phenomena encountered in such a structure, we will consider, subsequently, two planes A and B perpendicular to the direction De, these planes delimiting an area I to the left of A, which is that of the entry guide, zone II, between A and B, which is that of the junction proper, and zone III, to the right of B, which is that of the outlet guides. In addition, one will note C the edge where the two exit guides meet.

Such a junction finds numerous applications in integrated optics, in particular in the production of so-called Mach-Zehnder modulators.

The properties of these devices are described, for example, in the article by ANDERSON I. entitled "Transmission performance of Y-junctions in planar dielectric waveguides", published in the review "IEE Proceedings on Microwaves Optics and Acoustics", 2, 1 pages 7-12 of January 1978, as well as in the ar BAETS R. and LAGASSE PE, titled "Calculation of Radiation Loss in Integrated Optics Tapers and Y-junctions" published in the journal Applied Optics 21, 11, pages 1972-1978 of June 1982.

• a) de la conversion de mode dans la région de la jonction (région II),
• b) de l'apparition d'un rayonnement dans l'angle de la jonction (zone C),
• c) du couplage entre les deux guides de sortie (région III).

This type of device raises three types of difficulty which must be understood in order to grasp the interest of the invention which will be described. It's about :

• a) mode conversion in the region of the junction (region II),
• b) the appearance of radiation in the angle of the junction (zone C),
• c) coupling between the two outlet guides (region III).

Regarding mode conversion, the phenomenon is as follows. The input guide 10 and the output guides 11 and 12 are designed to operate in single mode. This means that, for the operating wavelength used, the width and the thickness of the guides are such that only the fundamental propagation mode can be established. At a given thickness, the tolerance on the width of the guide is very small and if this width is excessive, higher order modes will be able to propagate. This is precisely what happens in zone II of the device where the junction 13 has a flared shape whose width is constantly increasing between plane A and plane B. The condition for maintaining the lowest mode is therefore no longer respected in this area and higher order modes may appear. Another phenomenon combines with the previous one to amplify this mode conversion: it is the diffraction of light in plane A. Indeed, in this plane, the guide presents a discontinuity. The propagation vector directed according to De in zone I, will present, in the plane A, an angular dispersion so that, after A, the propagation vector will no longer be directed along De in the entire plane of cross section. On the edges, this vector will be directed obliquely.

These two phenomena combine to break the monomode character of the structure and cause the mode conversion. As the output guides 11 and 12 are, by construction, single-mode there will be incompatibility between the front of the multimode wave which reaches these guides in the plane B and the front of the single-mode wave capable of propagating in the two guides Release. Part of the input light energy will then be dispersed in the region of plane B.

The second difficulty encountered in junctions of this kind is the appearance of radiation in the angle C. The quasi-spherical front of the wave which propagates from plane A towards plane B, in region II, collides edge C where a diffraction wave is formed having said edge at its center. This wave radiates throughout the device, including towards the entry guide. Only a small part of this diffracted wave satisfies the conditions allowing its propagation in the outlet guides; the rest are released and lost.

The last difficulty concerns the proximity of the exit guides in region III. This proximity has the effect of coupling the guides to each other, which causes (as in a directional coupler where this effect is used) the transfer of energy from one guide to the other. This transfer presents a quasi-periodicity when one moves away from the plane B. This effect is all the more marked when the value of the angular opening (2a) is lower. However, in the Y junctions, this angle is necessarily small (less than a few degrees) if we want to obtain a good transfer energy from the inlet guide to the outlet guides. This coupling is therefore important.

The object of the invention is precisely to overcome all these difficulties. This object is achieved thanks to the use of a diffraction grating arranged between the entry guide and the junction, this grating having a pitch which defines only two diffraction directions of orders different from zero, +1 and - respectively 1, this step being also chosen so that these two diffraction directions coincide with the directions of the two outlet guides.

According to a preferred variant, the structure comprises a third single-mode output guide, located between the first two and having the same direction as that of the input guide, this third guide corresponding to the order O of diffraction.

• - la figure 1, déjà décrite, représente une jonction Y selon l'art antérieur,
• - la figure 2 illustre un exemple de réalisation d'une structure de guidage conforme à l'invention,
• - la figure 3 représente la répartition énergétique dans les modes d'ordre O et 1,
• - la figure 4 représente schématiquement une coupe de la structure utilisant un réseau d'indice,
• - la figure 5 représente schématiquement une coupe de la structure utilisant un réseau à ondulation,
• - la figure 6 représente une variante dans laquelle on effectue une modulation électrooptique,
• - la figure 7 illustre les différentes étapes d'un procédé de fabrication d'un motif permettant la réalisation d'un réseau,
• - la figure 8 illustre le mécanisme de diffusion (a) et le profil de diffusion obtenu (b),
• - la figure 9 illustre le mécanisme d'échange ionique (a) et le profil obtenu (b),
• - la figure 10 montre les différentes étapes d'un procédé d'obtention d'un guide à réseau par interférométrie.

The characteristics of the invention will appear better after the description which follows, of an exemplary embodiment given by way of explanation and in no way limiting. This description refers to attached drawings in which:

• FIG. 1, already described, represents a Y junction according to the prior art,
• FIG. 2 illustrates an exemplary embodiment of a guide structure according to the invention,
• FIG. 3 represents the energy distribution in the modes of order O and 1,
• FIG. 4 schematically represents a section of the structure using an index network,
• FIG. 5 schematically represents a section of the structure using a corrugated network,
• FIG. 6 represents a variant in which electro-optical modulation is carried out,
• FIG. 7 illustrates the different stages of a method of manufacturing a pattern allowing the creation of a network,
• FIG. 8 illustrates the diffusion mechanism (a) and the diffusion profile obtained (b),
• FIG. 9 illustrates the ion exchange mechanism (a) and the profile obtained (b),
• - Figure 10 shows the different steps of a method of obtaining a network guide by interferometry.

As shown in FIG. 2, the structure of the invention comprises a single-mode input guide 20, of direction De, two symmetrical single-mode output guides 21 and 22, of directions Ds ₁ and ^D S ₂ inclined by +9 and -9 relative to the direction De, and a third outlet guide 23, located in the extension of the inlet guide, that is to say having a direction Ds ₃ coincides with the direction De. The junction between the guide entry and exit guides is via a flared area 25. The structure also includes a diffraction grating 24, located at the end of the guide 20 and at the entrance to junction 25. This grating has a pitch p.

n étant l'indice de réfraction du matériau.It is generally known that when an array of steps p is crossed by an optical radiation of wavelength λ, there appear diffracted beams of various orders m (m positive or negative integer) in directions ϕ defined by :
sin ϕ =

n being the refractive index of the material.

supé- rieure à 1, pour laquelle il n'existe pas de valeur réelle de l'angle ϕ.In the case of the invention, a network is used for which only the +1 and -1 orders exist, excluding orders greater than 1. This condition is achieved by making the quantity λ np at least equal to 0, 5. Indeed, for such a value, everything order m greater than 1 will lead to a quantity

greater than 1, for which there is no real value for the angle ϕ.

; pour À = 1,55µm et n=3.3 par exemple, le pas sera de l'ordre du micron.For a given wavelength, we will therefore choose a step p less than

; for À = 1.55µm and n = 3.3 for example, the pitch will be of the order of a micron.

.The directions of the beams of order +1 and -1 then make angles + ϕ and - ϕ with the normal, the angle ϕ being defined by:
sin ϕ =

.

The angle θ of the directions Ds ₁ and DS ₂ with the direction De and the pitch p of the network are then chosen so that ϕ is equal to θ. As for the beam of order 0, its direction is that of the incident beam.

In other words, an incident beam passing through the input guide 20 gives rise, after crossing the network 24, to three beams and only three: a +1 order beam passing through the guide 21, an order beam -1 borrowing the guide 22, and a beam of order 0 borrowing the guide 23.

(x) où J₁ est la fonction de Bessel d'ordre 1 et où x est un paramètre fonction de la longueur L du réseau. Pour un réseau d'indice, où l'indice varie périodiquement d'une quantité An, le paramètre x est égal à

. L'énergie correspondant à l'ordre 0 est, elle, proportionnelle à

(x) où J_O est la fonction de Bessel d'ordre 0. Selon les valeurs que l'on donne à la longueur L du réseau, donc au paramètre x, on peut répartir la proportion d'énergie dans les guides dans certaines limites.With regard to the energy distribution in the different output guides, it will be observed that, conventionally, the energy diffracted by a network in the ± 1 orders is proportional to

(x) where J ₁ is the Bessel function of order 1 and where x is a parameter depending on the length L of the network. For an index network, where the index varies periodically by an amount An, the parameter x is equal to

. The energy corresponding to order 0 is proportional to

(x) where J _O is the Bessel function of order 0. According to the values which one gives to the length L of the network, therefore with the parameter x, one can distribute the proportion of energy in the guides within certain limits .

(x) et

(x). Le point de fonctionnement M correspond à un maximum d'énergie dans les ordres ±1, l'énergie dans le mode 0 n'étant toutefois pas nulle. Le point N correspond à une égalité d'énergie dans les trois modes. Le point P, pour lequel J'(x) = 0, correspond à l'absence d'énergie dans l'ordre 0, auquel cas on peut se passer du troisième guide 23, l'énergie diffractée étant uniquement concentrée dans les deux guides 21 et 22. Mais en général, on utilisera ce troisième guide, qui peut servir à des fins diverses, par exemple à une détection, à une contre réaction, etc...This question can be clarified using in Figure 3 where the two curves are schematically drawn

(x) and

(x). The operating point M corresponds to a maximum of energy in the ± 1 orders, the energy in mode 0 however not being zero. The point N corresponds to an energy equality in the three modes. The point P, for which J '(x) = 0, corresponds to the absence of energy in order 0, in which case we can do without the third guide 23, the diffracted energy being only concentrated in the two guides 21 and 22. But in general, we will use this third guide, which can be used for various purposes, for example for detection, feedback, etc ...

Figures 4 and 5 schematically show the layout of the network in the case of an index network (Figure 4) and in that of a ripple network, also called corrugation (Figure 5).

The index network (FIG. 4) is formed by slices of indices n, and n ₂ which alternate periodically. The variations in index can be obtained by diffusion as will be seen below.

The wavy network (FIG. 5) comprises a corrugated structure 72 which has the effect of periodically varying the effective thickness of the optical guide 40. It is known that the effective thickness of a wave guide is equal to the sum of three terms: the geometric thickness of the guide, the depth at which, in the substrate, the amplitude of the guided wave falls to 1 / e of its value and the distance at which, in the superstrate, the amplitude of l guided wave drops to 1 / e of its value. With a structure such as that of FIG. 5, where the thickness of the optical guide varies periodically, the effective thickness of the guide therefore also varies periodically. This periodicity in the propagation conditions creates the diffraction grating.

In a ripple network, its effect is limited to the region of the guide located near the corrugated surface and does not extend to the entire guide. In an index network (Figure 4), zones with different indices can affect the entire guide. As a result, the efficiency of a corrugated network is generally lower than that of an index network. This weakness will be compensated for by giving the network a greater length.

The advantages of the invention are clear. Thanks to the presence of the diffraction grating, the wave vector of the incident radiation, vector which dictates the propagation of the wave, is modified into wave vectors adapted to the directions of the output guides. Optimization of energy transfer is thus obtained, which was not the case in the prior art. Therefore, the diffraction losses in the corners D and E are reduced.

Regarding the mode conversion phenomenon, we can observe this. The network used has a low resolution in wavelength, because it includes a small number of "lines". Indeed, as has been explained above, the pitch of the grating is of the order of a micrometer, while the width of the waveguide is of the order of 8 to 10 μm. The network therefore does not include more than a dozen lines. The low resolution which results therefrom means that the energy of a high order mode, as it results from the conversion phenomenon mentioned above, will be diffracted towards the two guides 21 and 22 with a very small angular deviation with respect to in the fundamental mode. In this way, the losses due to the mode conversion are reduced compared to the prior art.

In addition, it is always possible to chan manage the shape of the network to improve energy distribution. We can for example give different areas of the network different lengths, for example short on the edges of the guide and large in the center.

Finally, the angular opening 28 can be large (for example several tens of degrees), which considerably reduces the coupling phenomena between outlet guides. The drawbacks mentioned above are therefore well avoided.

The structure which has just been described may consist of materials with a high electro-optical coefficient (LiNbO ₃ , InP, GaAs, etc.). One can then have the electrodes in the vicinity of the junction, to obtain a modulation of the light. The phenomenon implemented is a little different from that of conventional modulators. It can be explained with the help of FIG. 3, already described and of FIG. 6 which represents a network of index 24 inserted between two electrodes 30, 31 the first is connected to ground and the second is brought to a voltage V adjustable. The electric field applied to the electro-optical material constituting the network modifies the indices n _I and n ₂ of the different zones of the network, therefore the difference An, and consequently the parameter x mentioned above. If the applied voltage V is modulated between two values V ₁ and V ₂ , this is equivalent to a variation of x between two values x ₁ and x ₂ . If we refer to Figure 3, we see that the energy diffracted in the +1 and -1 orders then varies between two values E ₁ and E ₂ . A modulation of the light is thus obtained in the two output guides 21 and 22 (and of course in the guide 23).

• - répartition de l'énergie : 34% dans les deux guides extrêmes 21 et 22 et 8% dans le guide central 23, avec x=1,9, ce qui correspond sensiblement au point M de la figure 3,
• - répartition égale dans les trois guides de sortie (point N) avec x de l'ordre de 1,4 à 1,5,
• - avec un guide de Ti:LiNb0₃ de largeur 8 um, θ = 30° à la longueur d'onde de 1,55 µm; le pas p du réseau est de 1,4 µm et la longueur L est égale à 78 µm; le paramètre x vaut 1,9 ;
• - pour θ = 45°, λ = 1,55 µm, p = 1,0 µm et L = 60 µm, le paramètre x est égal à 1,45.

As an example, the following few numerical values can be indicated:

• energy distribution: 34% in the two extreme guides 21 and 22 and 8% in the central guide 23, with x = 1.9, which corresponds substantially to point M in FIG. 3,
• - equal distribution in the three outlet guides (point N) with x of the order of 1.4 to 1.5,
• - with a Ti guide: LiNb0 ₃ with a width of 8 μm, θ = 30 ° at the wavelength of 1.55 μm; the network pitch p is 1.4 µm and the length L is equal to 78 µm; the parameter x is 1.9;
• - for θ = 45 °, λ = 1.55 µm, p = 1.0 µm and L = 60 µm, the parameter x is equal to 1.45.

With regard to the manufacture of the device which has just been described, known techniques can be used such as those which are described in the article by PUN EYB and YI-YAN Alfredo entitled "Fabrication of Periodic Waveguides by Ion Exchanging "published in the journal" Applied Physics Letters ", 38, 9, p. 873-874 (May 1981) and in the article by PUN EYB, WONG KK, ANDONOVIC I., LAYBOURN PJR and DE LA RUE RM entitled "Efficient _W ave Guide Bragg-Deflection Grating on LiNbO ₃ " published in the journal Electronics Letters 18, 17, p. 740-742 (August 1982).

Figures 7 to 10 show some possible manufacturing process steps.

In FIG. 7, firstly, a layer 41 of doping material (FIG. A) is then deposited on a substrate 40, then a layer 42 of photoresist (FIG. B) which is exposed (FIG. C) through a mask 43 pierced openings 44. After development, there remain strips 45 of resin separating openings 46 (FIG. d). Through these openings a chemical or ionic attack on layer 41 is carried out. Then we remove the strips of resin 45, which leaves sub sister strips of dopant 47 on the substrate 40 (Figure e).

Another method which leads to the same result consists in depositing on the substrate 40 a layer 51 of photoresist (FIG. F), to expose this layer through a mask 52 pierced with openings 53 (FIG. G). After development, there remain strips 54 of resin (FIG. H). Next, a metal layer 55 is deposited (FIG. I) then, by removal ("lift-off"), the parts of this metal layer which cover the strips of resin are made to disappear. The pattern in Figure e is then obtained.

Having obtained this pattern either by one way or by the other, the process continues as shown in FIG. 8. The strips 47 then consist of a doping material, which diffuses into the substrate 40. This diffusion is essentially transverse but also has a side part. This distribution is shown schematically by the arrows in Figure a. After diffusion, a guide layer 50 of continuous doping is obtained as illustrated in FIG. B.

Instead of using the diffusion of a dopant, an ion exchange process can be used, as indicated in FIG. 9. The zones 47 are then formed by a mask which covers the substrate. When brought into contact with an ionic solution 54, this structure is the site of an ion exchange. For example Na ⁺ ions (56) from the glass substrate are exchanged with Ag ⁺⁺ ions (58) from the solution. This exchange is preferably carried out in a transverse direction, but also in a lateral direction (arrows in part a). As soon as the ion exchange is finished, the mask 47 is removed. The guide layer 60 finally obtained is still time and is represented on part b. EDe is similar to that of FIG. 8b, except that the effect obtained is now maximum with regard to the openings of the mask whereas, in the previous case, the diffusion was minimum . In both cases, the guide layer (50 or 60), on its substrate 40, sees its index vary periodically.

Figure 10 illustrates a final manufacturing process. On a substrate 40 coated with a guide layer 64 (FIG. A) a layer 66 of photoresist is deposited (FIG. B). This is exposed by two coherent rays 67 and 68 which interfere with the photoresist (figure c). The interference fringes thus obtained make it possible, after development, to obtain a layer of photoresist 70 of periodic thickness (FIG. D). By ion machining, through this layer (FIG. E), an undulating guide layer 72 is obtained overlying the substrate 40.

Claims (5)

1. Optical guide structure comprising a single-mode input guide (20) having a certain direction (De), two single-mode output guides (21, 22) having directions (Ds _l , Ds ₂ ) inclined symmetrically with respect to to the direction (De) of the entry guide and a flared junction zone (25) between the entry guide and the exit guides, characterized in that it further comprises a diffraction grating (24) arranged between the entry guide and the junction, this grating having a pitch which defines only two diffraction directions of orders different from zero, +1 and -1 respectively, this pitch being chosen so that these two diffraction directions coincide with the directions of the first two exit guides. 2. Structure according to claim 1, characterized in that it comprises a third single-mode output guide (23) located between the first two and having the same direction as that of the input guide. 3. Structure according to claim 1, characterized in that the network is surrounded by two electrodes (30, 31) brought to different potentials. 4. Structure according to claim 1, characterized in that the network (24) is an index network. 5. Structure according to claim 1, characterized in that the network (24) is obtained by undulating the thickness of the guide layer.

Images